Exploring Data Visualization with Scientific Results
Misha Ash
2018-06-03
introduction
Exploring conceptualization through the lens of spontaneous co-thought gestures produced during abstract relational reasoning.
“We are more apt to make a grasping gesture when we speak of grasping an idea than when we speak of grasping a doorknob.” Benjamin Lee Whorf 1
People rountinely gesture, both in conversation with others and when thinking by themselves.2
Gestures can be metaphorical, representing non-spatial attributes with spatial attributes, and some are analogical, representing non-spatial relations with spatial relations.3
Do people gesture even when thinking about abstract, non-spatial relations?
What factors influence the production of analogical co-thought gestures?
Is it worth visualizing the basic structure of experimental design?
DESIGN
Within-subjects (2 x 2 x 3)
- spatiality of relations
- spatial
- non-spatial
- axis of relations
- vertical
- horizontal
- problem difficulty
- hard
- easy
- invalid
- spatiality of relations
(Blocked by problem type and axis with difficulty pseudorandomly interleaved.)
Visualizations
- The factors of an experiment are typically described without visualization, particularly when there are more than two factors. When they are visualized, it is usually as a matrix grid.
- Alternative visual forms may more readily capture the hierarchical structure of crossed multifactorial designs.
- Thoughts about visualizing this structure? Is it helpful?
- Could also try a tree map with groups and subgroups.
- Need colors and labelling.
Linear syllogisms were used to systematically vary relation type and difficulty.
METHODS
After the experimenter left the room, participants solved 40 linear syllogisms, blocked by trial type (n = 24).
Each syllogism was presented for 10 seconds.
Participants solved each problem by saying “yes” if true, “no” if false, and “can’t tell” if invalid.
Data of participants was coded for response accuracy, gesture presense, and axis of gesturing.
Waffleplots not rendering (shown as image).
Typical visualization of mean comparison results.
Spatial trials were more likely to elicit gesture (p = .002).
Trial axis did not affect gesture production (p = .14; not shown).
A limitation of barplots.
As Page Piccinini and others have pointed out, barplots can hide the distribution of data!
Page suggests we abandon barplots in favor of boxplots.
Histograms
Boxplots
Boxplot more clearly shows the distribution.
Datapoints seem unnecessary in this case: they do not further reveal the data and overlap at the extremes.
Using split violin plots to visualize distributions of mean comparisons.
- Not sure how kernel density estimation (KDE) is affecting this visualization, but the distribution extends beyond the range of the data.
Order effects (chronological)
More gestures were produced when spatial relations are presented before abstract relations.
However, more gestures were produced in the spatial condition even when non-spatial relations were presented first.
SKETCHES (temp)
library(igraph)
library(ggraph)
dendrogram <- as.dendrogram(hclust(dist(iris[, 1:4])))
ggraph(dendrogram, 'dendrogram') +
geom_edge_elbow()
ggraph(dendrogram, 'dendrogram', circular = TRUE) +
geom_edge_elbow() +
coord_fixed()
dendro_data <- level_01_data %>%
select(SPACE_IV, AXIS_IV, DIFF_IV)
dendro_study_design <- as.dendrogram(hclust(dist(dendro_data)))
ggraph(dendro_study_design, 'dendrogram') +
geom_edge_elbow()
ggraph(dendro_study_design, 'dendrogram', circular = TRUE) +
geom_edge_elbow() +
coord_fixed()Whorf, B. L. (1944) The Relation of Habitual Thought and Behavior to Language. Etc: A Review of General Semantics, 1(4), 197–215.↩
e.g., Kita, S., Alibali, M. W., & Chu, M. (2017). How do gestures influence thinking and speaking? The gesture-for-conceptualization hypothesis. Psychological Review, 124(3), 245–266.↩
Cooperrider, K., & Goldin-Meadow, S. (2017). When Gesture Becomes Analogy. Topics in Cognitive Science, 9(3), 719–737.↩